Measurements of the free luminal ER Ca(2+) concentration with targeted "cameleon" fluorescent proteins

(1)

Article

Reference

Measurements of the free luminal ER Ca(2+) concentration with targeted "cameleon" fluorescent proteins

DEMAUREX, Nicolas, FRIEDEN, Maud

Abstract

The free ER Ca(2+) concentration, [Ca(2+)](ER), is a key parameter that determines both the spatio-temporal pattern of Ca(2+) signals as well as the activity of ER-resident enzymes.

Obtaining accurate, time-resolved measurements of the Ca(2+) activity within the ER is thus critical for our understanding of cell signaling. Such measurements, however, are particularly challenging given the highly dynamic nature of Ca(2+) signals, the complex architecture of the ER, and the difficulty of addressing probes specifically into the ER lumen. Prompted by these challenges, a number of ingenious approaches have been developed over the last years to measure ER Ca(2+) by optical means. The two main strategies used to date are Ca(2+)-sensitive synthetic dyes trapped into organelles and genetically encoded probes, based either on the photoprotein aequorin or on the green fluorescent protein (GFP). The GFP-based Ca(2+) indicators comprise the camgaroo and pericam probes based on a circularly permutated GFP, and the cameleon probes, which rely on the fluorescence resonance energy transfer (FRET) between two GFP mutants of different [...]

DEMAUREX, Nicolas, FRIEDEN, Maud. Measurements of the free luminal ER Ca(2+)

concentration with targeted "cameleon" fluorescent proteins. Cell Calcium , 2003, vol. 34, no. 2, p. 109-19

DOI : 10.1016/S0143-4160(03)00081-2 PMID : 12810053

Available at:

http://archive-ouverte.unige.ch/unige:30406

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Cell Calcium 34 (2003) 109–119

Invited review

Measurements of the free luminal ER Ca ² ⁺ concentration with targeted “cameleon” fluorescent proteins ^夽

Nicolas Demaurex

^∗

, Maud Frieden

Department of Physiology, University of Geneva Medical Center, 1, Michel-Servet, CH-1211 Geneva 4, Switzerland Received 3 January 2003; received in revised form 19 March 2003; accepted 21 March 2003

Abstract

The free ER Ca²⁺concentration, [Ca²⁺]_ER, is a key parameter that determines both the spatio-temporal pattern of Ca²⁺signals as well as the activity of ER-resident enzymes. Obtaining accurate, time-resolved measurements of the Ca²⁺activity within the ER is thus critical for our understanding of cell signaling. Such measurements, however, are particularly challenging given the highly dynamic nature of Ca²⁺signals, the complex architecture of the ER, and the difficulty of addressing probes specifically into the ER lumen. Prompted by these challenges, a number of ingenious approaches have been developed over the last years to measure ER Ca²⁺by optical means. The two main strategies used to date are Ca²⁺-sensitive synthetic dyes trapped into organelles and genetically encoded probes, based either on the photoprotein aequorin or on the green fluorescent protein (GFP). The GFP-based Ca²⁺indicators comprise the camgaroo and pericam probes based on a circularly permutated GFP, and the cameleon probes, which rely on the fluorescence resonance energy transfer (FRET) between two GFP mutants of different colors. Each approach offers unique advantages and suffers from specific drawbacks. In this review, we will discuss the advantages and pitfalls of using the genetically encoded “cameleon” Ca²⁺indicators for ER Ca²⁺measurements.

Keywords: Cameleon; Pericam; Camgaroo

1. Introduction

Measuring the dynamic Ca²⁺signals occurring within the lumen of intracellular organelles poses a number of specific constraints. To obtain reliable estimates of the free ER Ca²⁺ concentration, one should be able to label the organelle se- lectively and to detect Ca²⁺ fluctuations that are orders of magnitudes higher than in the cytosol. At the same time, the function of the cell and of its organelles should be preserved.

Not surprisingly, all the approaches developed to date rely on optical techniques, which allow the measurement of labeled structures located deep inside cells or tissue with high sensitivity and minimal interference. The first attempts relied on synthetic fluorescent chelators such as Mag-fura-2, which were found to accumulate into the ER under high temperature loading conditions[1–3]. The dyes are easy to load and provide a bright fluorescence signal, but the targeting is not specific and excess cytosolic dye has to be removed by per- meabilization or by dialysis via a patch pipette to reveal the

夽Supplementary data associated with this article can be found at doi:10.1016/S0143-4160(03)00081-2.

∗Corresponding author. Tel.:+41-22-702-51-11.

E-mail address: nicolas.demaurex@medecine.unige.ch (N. Demaurex).

ER staining. Furthermore, the dyes do not accumulate well in the ER of some cell types and gradually leak out from cells at physiological temperatures. The use of the biolu- minescent protein aequorin provided a significant improvement, as this genetically encoded Ca²⁺sensor allows both stable and specific ER labeling[4–8]. However, aequorin requires the incorporation of the cofactor coelenterazine and is very difficult to image because it emits less than one photon per molecule. Furthermore, aequorin is irreversibly consumed by Ca²⁺, precluding long-term measurements in Ca²⁺-rich compartments such as the ER. Chimeras proteins based on the green fluorescent protein (GFP) and calmodulin (CaM) combine the benefits of molecular targeting, tuneable Ca²⁺affinity, and bright fluorescence[9,10]. Much effort has been devoted to the molecular engineering of these promis- ing Ca²⁺indicators. Several types of GFP/CaM probes have (and are still being) developed, in which Ca²⁺ either alters the fluorescence of a circularly permutated GFP (camgaroos, G-Cam, and pericams) or promotes the reversible association of two GFP mutants of different colours (cameleons). This last approach takes advantage of the radiation-free transfer of energy that occurs when two fluorophores with overlapping excitation/emission spectra are brought in close proximity (1–8 nm). By linking together the two GFP partners with a

doi:10.1016/S0143-4160(03)00081-2

(3)

Fig. 1. Structure of cameleon probes. The probes use FRET between GFP partners to monitor the Ca²⁺-dependent binding of calmodulin to a target peptide (M13 or CKKp). In the yellow cameleon shown here, binding of Ca²⁺ to the high and low affinity sites of calmodulin induces FRET between cyan and yellow GFP partners. The reversible Ca²⁺-dependent changes in YC conformation are thus visible as changes in colour.

flexible linker sequence, intramolecular fluorescence resonance energy transfer (FRET) can be reversibly induced, and conformation changes within the fusion protein are trans- duced into changes in colour (Fig. 1).

Romoser et al. were the first to use this approach to generate a Ca²⁺probe, by inserting the calmodulin-binding do- main from smooth muscle light chain kinase (M13) between two GFPs[11]. Binding of Ca²⁺-bound CaM to this hybrid protein increases the distance between the two GFP mutants, resulting in a decrease in FRET. However, the probe reported the changes in Ca²⁺/CaM activity, rather than in the free Ca²⁺concentration. Miyawaki et al. went a step further by designing the cameleons, which contain both CaM and its binding peptide M13 as a Ca²⁺sensing module sandwiched between two GFP[9]. Upon binding of Ca²⁺, CaM prefer- entially wraps around its neighboring M13 peptide, bringing the two GFP together and increasing FRET. As both CaM and M13 are included within the cameleon protein, the M13 peptide is much less likely to interact with endogenous CaM proteins and to interfere with cellular functions. The original cameleons used a blue fluorescent protein (BFP) as a FRET donor and GFP as the acceptor. However, BFP has a poor quantum yield, is very sensitive to bleaching and has its peak excitation in the UV range, raising concerns of phototoxicity and autofluorescence. The more efficient yellow cameleons (YC) use the enhanced cyan emitting mutant of the green fluorescent protein (ECFP) as the donor and the yellow-emitting mutant (EYFP) as the acceptor (Fig. 1).

ECFP and EYFP are spectrally well separated with sufficient overlap between their respective emission and excitation peaks to generate a good FRET signal. These two GFP mutants are now the preferred FRET partners for biological applications. The increase in FRET efficiency can be easily detected by exciting the ECFP donor at 430 nm and measur-

ing the shift in fluorescence emission from 480 nm (the peak of the ECFP donor emission) to 535 nm (the peak of the EYFP acceptor emission). In the YC, the shift from 480 to 535 nm is directly proportional to the amount of Ca²⁺linked to YC and the probes behave as ratiometric Ca²⁺indicators.

The cameleons immediately generated tremendous interest, because the use of GFP as a Ca²⁺ indicator provides several key advantages for in vivo imaging. The protein can be expressed in a variety of cells and spontaneously generates a bright fluorescence without the need for a cofactor.

GFP can be targeted precisely to specific tissues, organelles, or cellular micro domains by simple molecular engineering.

This unique property has been exploited to measure Ca²⁺ signals with YC in cells or organisms previously unreach- able by conventional techniques (Table 1). The ability to ad- just the calcium affinity of YC by mutating calcium-binding acidic residues within its calmodulin module further allow to design probes that match the calcium concentration within the organelle of interest. Unfortunately however, the original probes displayed a marked pH dependency, suffered from a small dynamic range, and were difficult to express in some cells or organelles because the folding and maturation of the protein at 37^◦C was poor. Since then, substantial work has been devoted to improve these GFP-based Ca²⁺indicators.

The published steps retracing their ongoing evolution toward an optimal Ca²⁺probe are reviewed in the next section.

2. Design and evolution of cameleons

The successive rounds of cameleon engineering are reflected in the nomenclature, with probes numbered according to their decreasing Ca²⁺ affinities and successive version (Table 1). The first series comprised three probes with distinct Ca²⁺affinities: YC2, bearing the “wild-type”

CaM module (E31/E104), YC3, bearing the E104Q substitution (E31/Q104) and YC4, bearing the E31Q substitution (Q31/E104). YC2 displays a biphasic Ca²⁺-binding curve and reports [Ca²⁺] variations between 100 nM and 10␮M. YC3 has a slightly lower affinity for Ca²⁺ and a monophasic response, and is suitable for monitoring [Ca²⁺] changes from 1 to 100␮M. YC4 has a much lower affinity and reports [Ca²⁺] changes in the range 10␮M to 1 mM [9]. This first series of YC has been used extensively for Ca²⁺ measurements in the ER, with the lowest affinity probe (YC4ER) matching the ER concentrations found in resting cells (∼500␮M) and the higher affinity probe (YC3_ER) matching the values found in Ca²⁺-depleted cells (Table 1).

The original YC suffered from the strong pH-dependency of the EYFP mutant, whose fluorescence is quenched by acidification around the physiological pH (pK_a6.9). Because the yellow fluorescence is more pH sensitive than the cyan fluorescence, a drop in pH thereby mimics a drop in [Ca²⁺] during ratio measurements. The pH changes can be recog- nized from the Ca²⁺changes, however, as both the cyan and

(4)

N.Demaurex,M.Frieden/CellCalcium34(2003)109–119111 Table 1

Properties and applications of cameleon probes

Series pH dependency (pKa) Maturation (h)^a Probe Ca²⁺range (␮M) Dynamic range^b Targeting Cells References

Yellow cameleon 6.9 24–72 YC2 0.1–10 1.6 Cytosol HeLa [9,12]

HEK-293 [15,26,35,42]

Drosophila melanogaster (transgenic fly) [29]

Nucleus HeLa [9]

S. pombe (yeast) [43]

Secretory vesicle PC-12, MIN6 [44]

Mitochondria Neonatal rat cardiomyocytes [27]

HeLa [12]

YC3 0.5–100 1.7 ER HeLa [9]

Xenopus leavis oocytes [16]

HEK-293 [15,45]

YC4 10–1000 1.4 ER HeLa [9,12]

X. leavis oocytes [16]

HEK-293 [15,26,35,42]

Mouse embryonic fibroblasts [15,26,35,42]

Rat embryonic R6 fibroblasts [15,26,35,42]

MIN6 [17]

Human endothelial cells EA.hy926 [39]

Improved cameleon 6.1 24–72 YC2.1 0.1–10 1.6 Cytosol HeLa [10,31]

Rat hippocampal neuron [10]

Arabidopsis thaliana (plant cells) [46,47]

Caenorhabditis elegans (transgenic worm) [28]

Pancreatic islets [32]

Mitochondria HEK-293 [45]

Caveolae Bovine aortic endothelial cells [30]

YC3.1 0.5–100 1.7 Cytosol HeLa [10]

Nucleus HeLa [31]

Mitochondria HeLa [12]

YC4.1 10–1000 1.3 Mitochondria HeLa [12]

HEK-293 [15,26,35,42]

Citrine cameleon 5.7 24–120 YC3.3 0.5–100 1.7 Golgi HeLa [14]

CKKp cameleon 6.1 48–120 YC6.1 0.5–1 2.1 Cytosol HeLa [18]

Rat hippocampal neuron

Venus cameleon 6.0 4 YC2.12 0.1–10 N.A. Cytosol Mouse Purkinje neuron [20]

Red cameleon 5.5 24–72 SapRC2 0.2–0.4 1.3 Cytosol HeLa, rat hippocampal neuron [48]

aTime lag between cell transfection and Ca²⁺ imaging.

bRmax/Rminratio measured in intact cells.

(5)

yellow fluorescence are decreased by the acidic pH, whereas Ca²⁺ produces opposite changes in the two wavelengths.

The built-in pH dependency of YC can thus be harnessed to monitor the ER pH which, fortunately, remains stable during cell stimulation with Ca²⁺mobilizing agonists[12]. How- ever, the establishment of precise [Ca²⁺]_ER values is difficult, because the pH of the ER may vary significantly from one cell to the other. The ER has a very high passive permeability for H⁺, and changes in cytosolic pH are readily trans- mitted to the ER[13]. Differences in the cell metabolic state are thus reflected as changes in both cytosolic and ER pH.

To account for this cell-to-cell variability, the maximal and minimal ratio values (Rmin and Rmax) must be determined independently for each cell during [Ca²⁺]ERmeasurements with YC.

To circumvent the pH dependency, the group of Roger Tsien added two new mutations within the EYFP module (V68L and Q69K), which decreased its pK_a to 6.1 [10].

These “improved cameleons” (YC2.1, YC3.1, and YC4.1, with Ca²⁺ sensitivities matching the YC2, YC3, and YC4 probes, respectively) are much less pH sensitive and can be readily targeted to the ER by adding the appropriate signal and retention sequences. However, folding of the modi- fied probes is much less efficient at 37^◦C and a prolonged incubation period at 30^◦C is often required to ensure protein expression in the ER. In addition, the dynamic range of the ER-targeted low affinity probe (YC4.1ER) was reduced (Rmax/Rmin: 1.3), impairing its ability to measure Ca²⁺accurately in the ER. Thus, despite their better pH stability, the improved YC have not been used extensively for ER Ca²⁺ measurements.

A further improvement in the pH sensitivity of EYFP was conferred by the Q69M or “citrine” mutation[14]. This mu- tation decreased the pK_a of EYFP to 5.7, removed its sensitivity to halides and improved the folding of the protein at 37^◦C as well as its photostability. Inclusion of citrine as the FRET acceptor in a cameleon probe resulted in a more pH-resistant cameleon, YC3.3, with Ca²⁺ affinity matching the YC3 probe. The ratio of the emitted fluorescence (528/476 nm) was stable in the pH range 6.5–7.5, whereas the other properties, which were assessed in vitro only for the cytosolic version, were similar to the previous cameleons (Table 1). A citrine-based YC fused with the N-terminus of the human galactosyltransferase type II (GT-YC3.3) was used to measure the Ca²⁺ signals in the Golgi complex, a compartment that is more acidic than the ER (pH ∼6.5).

The probe was near saturation in unstimulated cells, indicating that the Golgi Ca²⁺concentration is high, and Ca²⁺ changes elicited by ionophores could be readily detected despite the acidic conditions[14]. Because of their improved photo stability and reduced sensitivity to pH and chloride, the citrine-based cameleons should thus be more reliable than the original YC for Ca²⁺measurements in acidic compartments.

All the YC suffer from a small dynamic range, the maximal changes in ratio between the Ca²⁺-free and

Ca²⁺-saturated state (R_max/R_min) being around two in vitro, and ranging between 1.3 and 1.7 in situ (Table 1). In the ER, the changes in ratio observed during physiological stimulation cover only a fraction of the full dynamic range, less than 10%[12,15–17]. This limited range makes the quantification of the physiological fluctuations in [Ca²⁺]_ERvery difficult. To improve the dynamic range, the group of Mitsu Ikura designed a new cameleon bearing a Ca²⁺-sensing module derived from CKKp, the CaM-binding peptide of CaM-dependent kinase kinase [18]. Based on the structure of CKKp-bound calmodulin, which suggested that CKKp could fit within the linker region connecting the two EF-hand domains of CaM, CKKp was sandwiched between the N-terminal and the C-terminal fragment of CaM. The resulting Ca²⁺-sensing module (N-CaM-CKKp-C-CaM) is shorter than the original CaM-M13 module, decreasing the distance between the two GFP from 60 to 40 Å in the Ca²⁺-bound state. The closer proximity of the two CFP/YFP partners increased the FRET efficiency in the presence of Ca²⁺, and thus the dynamic range. This new CKKp-based YC, dubbed YC6.1, displayed a monophasic Ca²⁺-dependency with an apparent dissociation constant of 110 nM, and responded in a linear manner to Ca²⁺changes ranging from 50 nM to 1␮M. Comparison of the cytosolic version of YC6.1 with the YC2.1 probe in HeLa cells stimulated with histamine revealed that the new probe had a slower response time, but a much improved dynamic range (2.1 vs. 1.4). The YC6 series thus offers an interesting alternative to other YC which are based on the CaM-M13 module, especially when a good signal-to-noise ratio is critical, as is the case for [Ca²⁺]_ER measurements. Both, a nuclear and an ER-targeted version of the CKKp-based YC have been generated (YC6.1_nuand YC6.2_ER), but these probes remain to be evaluated.

The increase in dynamic range achieved with the YC6 probes was less than anticipated based on the shorter distance predicted by the structure-based design, possibly because the FRET efficiency also depends on the proper ori- entation between the donor and acceptor chromophores. The YC6 probes also used as FRET acceptor the second generation “improved” EYFP (V68L/Q69K), which still retains significant pH and halide sensitivity, is sensitive to bleaching, and matures very slowly at 37^◦C or in some organelles.

Because the FRET signal is mainly limited by maturation of the EYFP acceptor (see below), attempts were made to generate EYFP mutants that mature more efficiently. For this purpose, Nagai et al. introduced the mutation F46L, found during the random mutagenesis of the pericam probes to greatly enhance maturation at 37^◦C [19,20]. The F46L mutation greatly accelerated the rate of chromophore formation, an effect that was specific to EYFP and could not be transposed to the blue, cyan or green versions of the fluorescent protein. Inclusion of the well known folding mutations F64L/M153T/V163A/S175G further increased the folding of EYFP by about ∼15-fold, generating a probe that folds and re-oxidizes much faster than the original EYFP. The

(6)

N. Demaurex, M. Frieden / Cell Calcium 34 (2003) 109–119 113

faster folding and maturation resulted in a 30-fold increase in the fluorescence intensity when the protein was expressed in E. coli, and in an eightfold increase in fluorescence when the protein was expressed in mammalian cells. These bright probes, that fold and form the chromophore very efficiently at 37^◦C, were named “Venus”. The excitation and emission spectra of the Venus probes, as well as their extinction coefficient (the ability of the fluorophore to capture light) and fluorescence quantum yield (the ability of the fluorophore to convert incoming light into emitted fluorescence) were not significantly altered compared to the original EYFP.

The rate of bleaching was similar to the other YFPs, and was not improved by the addition of the citrine mutation (Q69M), but the chloride and pH sensitivity of the Venus fluorescence was reduced (pKa 6.0). The pH insensibility of the Venus probes allowed to image dense-core granules, which are very acidic. Importantly, all cells expressing dense-core targeted Venus displayed a similar fluorescence pattern, indicating that cell-to-cell variability of expression was improved in the Venus mutant. Using Venus as a FRET acceptor in place of the EYFP module, Nagai et al. produced a rapidly maturing cameleon probe, YC2.12 [20]. This

“Venus-cameleon” was tested in a mouse cerebella slice by ballistic particle-mediated gene transfer of the cDNA. After only 4 h, a variety of cells located throughout the slice were brightly fluorescent, with clean cytosolic staining observed in a Purkinje neuron. The YC2.12 probe responded ade- quately to Ca²⁺changes and its dynamic range appears to be similar to previous cameleons. Because of its fast and efficient maturation, the YC2.12 probe should allow the immediate detection of Ca²⁺ signals in freshly prepared cells or tissues.

In a different approach, a Red cameleon was constructed using the red fluorescent protein from Discosoma (DsRed) as FRET acceptor. The broad excitation spectra of DsRed prevented the use of CFP or YFP as FRET donors, but a good separation could be achieved by using a Sapphire fluorescent protein as donor. The Sapphire-Red cameleon probe (SapRC2) can be excited at 400 nm, and the emission ratio measured at 580/510 nm. The dynamic range of SapRC2 expressed in HeLa cells was poor, around 1.3, and significant bleaching of the DsRed was reported when using a laser scanning confocal microscope. Because both Sapphire and DsRed red are pH insensitive (pKa∼5.5), the SapRC probe is resistant to acidification down to pH 5.5. The response time to histamine stimulation of HeLa cells, however, was relatively slow. This might reflect the tendency of DsRed to aggregate, which led to the formation of homotetramers of SapRC2. Indeed, the fact that the homotetramers were still able to function as a Ca²⁺ indicators was surprising, given the large conformational change required to produce the Ca²⁺-dependent change in the FRET signal. The recent engineering of a monomeric DsRed[21]opens the way to the generation of monomeric red cameleons, which might display more interesting properties and offer an alternative to the yellow cameleons.

Each of the probes described above has specific advantages. A better YC construct could be generated by combining their most attractive properties, for instance by using the Ncam-CKKp-Ccam as Ca²⁺ sensing module and Venus as the yellow FRET acceptor. Such a construct should mature rapidly at 37^◦C and exhibit a bright fluorescence, with a purely monophasic Ca²⁺dependency and a dynamic range higher than 2. Better probes will certainly be generated by such simple modular engineering of existing cameleons or by structure-based design of fluorescent proteins.

3. Critical factors affecting cameleon measurements

The main advantage of genetically encoded indicators is that cells will produce the fluorescent protein endogenously.

The cell sorting machinery is then exploited to address to specific compartments the probes fused to specific targeting signals, retention sequences, or parts of adaptor proteins.

This is also the main disadvantage of the approach, which entirely relies on the cell-transcriptional, -translational, and post-translational machinery. This last step is critical, as the cells not only have to put the probe in the correct compartment, but should allow the proper maturation and folding of a fluorescent Ca²⁺-binding protein made up of two GFP mutants. Many steps between the addition of the probe cDNA and its expression as a functional Ca²⁺probe can be opti- mized, as discussed below, but remain dependent on a host of cell-specific factors that are only partially in the hands of the experimenter.

3.1. Expression

Measurements with YC critically depend on the ability of the recipient cells or tissue to take up the gene transferred and to transcribe and translate the genetic material efficiently. Regardless of the technique used for gene transfer, a significant time lag is required before Ca²⁺ measurements can be performed. For cDNA transfection with either Ca²⁺-phosphate or cationic lipids, the measurements are typically performed between 2 and 5 days after gene transfer into the cells [9,12,15–17]. Attempts to maximize the efficacy of transfection often cause a significant fraction of cells to become brightly fluorescent, with most of the fluorescence originating from irrelevant cell locations. This indicates a defective maturation of the protein, and/or saturation of a step along its synthesis or trafficking pathway, leading to the accumulation of mature or immature forms of the probe into undesired cell compartments. Imaging measurements can still be performed on the subpopulation of cells with “correct” addressing, but these cells must be selected visually, which might bias the interpretation of the results. Fluorescence measurements in cell populations cannot be performed in this case, as they will be contami- nated by the strong signal of the mislabeled cells. One way around this problem is to transfect cells with low efficiency,

(7)

in order to have only a few cells with weak fluorescence, but correct staining. Another approach is to generate stable cell lines, which have the additional benefit to allow high-throughput fluorescence readouts in cell populations.

However, the generation of cell lines is tedious and relies on lengthy selection procedures. The use of lentiviral-based vectors provides a more efficient way to integrate the exogenous DNA into the host genome, even in cells failing to take up exogenous DNA applied with Ca²⁺-phosphate or lipid-based protocols [22]. This gene delivery system appears better suited for YC expression in cells or tissues.

However, the recipient cells must still be able to address the protein into the desired compartment where it must fold and mature properly to become fluorescent.

3.2. Maturation and folding

GFP maturation is a complex process that involves the folding of the protein into a native conformation, cycliza- tion of the internal tripeptide that forms the chromophore, and the subsequent oxidation of the chromophore, a process that requires molecular oxygen. All these steps are affected by the local environmental conditions existing in the different cell types and/or intracellular compartments where the probes are expressed. Differences in protein content, ionic strength, pH, and redox conditions will influ- ence GFP folding and chromophore maturation. Mutations within the GFP or in adjacent linker regions, which often alter the immediate environment of the chromophore, also affect GFP folding and maturation. As a result, the fluorescence properties of GFP-based probes often differ from one construct to the other, between different cell types, and between different subcellular locations. Among GFP mutants, the yellow-emitting mutant appears to be the most sensitive to structural or environmental alterations, probably because the aromatic substitution responsible for the∼20 nm shift to longer excitation and emission wavelengths (T203Y) increases the local polarizability immediately adjacent to the chromophore[23]. This renders the YFP fluorescence much more sensitive to photobleaching, protonation, and quench- ing by halides[24]. Because all cameleon probes except the recently developed red cameleons use EYFP as a FRET acceptor, the properties of the YC probes depend in large part on the properties of the EYFP module. Maturation of the EYFP also limits the FRET efficiency of the YC, because YC proteins bearing an immature EYFP acceptor are unable to engage in FRET. The light emitted by these immature YC thus dilutes the FRET signal of mature donor–acceptor pairs[25]. Thus, the efficiency of YC as Ca²⁺probes critically depends on the proper maturation and folding of their EYFP module.

GFP folding is a slow process that proceeds better below 37^◦C. Lowering the temperature has a beneficial effect on GFP folding, and the improper behaviour or targeting efficiency of some YC can thus be rescued by incubating the transfected cells at 30^◦C for a few hours[9]. GFP

maturation and targeting efficiency is also highly dependent on the protein expression levels. High levels of transfection favour the accumulation of immature proteins, which might either fail to generate fluorescence upon illumination (thereby reducing the quantum yield) or fail to be targeted appropriately. This affects the relevant signals in three ways: (1) mistargeting to compartments with stable, low Ca²⁺ values (silent compartments) decreases the dynamic range and bias the readout to lower values; (2) mistargeting to compartments with stable, but high Ca²⁺values (hot compartments) also decreases the dynamic range but bias the steady-state readouts to higher values; and (3) mistargeting to compartments exhibiting Ca²⁺transients renders the YC signal impossible to interpret. In addition, because each YC can bind four Ca²⁺, high concentrations of YC can significantly affect Ca²⁺ dynamics by their buffering power. Miyawaki et al. reported that whereas [Ca²⁺]_coscil- lations can be observed in cells expressing YC3.1 at concentration between 40 and 150␮M, these were never observed in cells with [YC3.1] > 300␮M [10]. Thus, the fluorescence properties of the YC probes depend on the construct and cell type used as well as on the organelle measured, and are affected by changes in protein expression levels, temperature, and redox conditions. A careful calibration is thus critical to obtain meaningful information using cameleon probes.

3.3. Technical limitations

Even with an ideal Ca²⁺ probe, capturing Ca²⁺ signals within the lumen of the ER poses a major technical chal- lenge. Ca²⁺signals are extremely versatile in both time and space, with highly local signals lasting a few milliseconds and global oscillatory responses lasting for hours or days.

This versatility is due in large part to the controlled release and uptake of Ca²⁺ from the ER, implying that the Ca²⁺ signals occurring within the ER lumen are equally versatile.

Imaging local and global Ca²⁺signals is already challenging in the cytosol, and becomes much more difficult in the ER because of the complex and dynamic nature of this organelle (Fig. 2). A high spatial resolution is required both in the axial and lateral direction to resolve the fine, convoluted structure of the ER. At the same time, a temporal resolution in the millisecond range is required to capture local Ca²⁺ events.

Rapid acquisition is also needed to “freeze” the moving ER structure in order to preserve the spatial resolution. For ratio measurements, images must be acquired at two different wavelengths and must present sufficient signal-to-noise to allow the quantification of the fluorescence signal. Fi- nally, several optical sections must be acquired to recon- struct the complex three-dimensional structure of the ER.

Thus, to capture Ca²⁺events in their totality, the entire vol- ume of the cell should be imaged at high-resolution in a few tens of milliseconds, at two wavelengths and with sufficient signal-to-noise, while using minimal illumination to avoid phototoxicity and reduce photobleaching.

(8)

Fig. 2. Dynamics of the ER in live HeLa cells. HeLa cells expressing ER-targeted GFP were imaged with a spinning-wheel confocal system (Visitech, Sunderland, UK¹), using 488 nm illumination and 535 nm emission. The Nipkow microscope was equipped with an AOTF-driven Argon ion laser and a CCD camera (Coolsnap HQ, Roper Scientific, Trenton, NJ) controlled by the MetaMorph software (Universal Imaging, West Chester, PA). Stacks of confocal images (300 ms exposure, 7 z sections, 0.5 um step) were acquired every 6 s for 4 min. Images were deconvolved with the Huygens algorithm (Scientific Volumetric Imaging, Ilversum, The Netherlands) using the Imaris software (Bitplane AG, Zurich, Switzerland). The middle plane of the confocal z stack is shown in this 4-min time-lapse movie to illustrate the highly dynamic nature of the ER. Bar: 2␮m.

Most [Ca²⁺]_ERrecordings published so far have not been performed under these ideal conditions, but instead using standard equipment consisting of a wide-field microscope equipped with a CCD camera and an emission filter changer to sequentially acquire cyan and yellow cameleon images. In this configuration, exposure times typically range between 0.5 and 2 s because cells with proper YC targeting are not very bright and illumination must be kept low to avoid pho- tochromism. The final time resolution thus rarely exceeds 1 ratio/s and is often closer to 1 ratio every 4–5 s. In these conditions, rapid Ca²⁺events are not detected and the spatial resolution is impaired by the mobility of the organelle.

Signal quantification is also problematic, because the divi- sion (ratio) of two separate images acquired at >1 s intervals generates significant amounts of noise on the ratio image.

The most significant problem occurs in regions of low fluorescence such as the organelle edges, whose displacement between sequential images leads to multiplication or divi- sion by near-zero values, producing “dark” or “hot” pixels on the ratio image. These irrelevant pixels can be disre- garded by applying a low intensity threshold or by merging

1We have no financial interests in any of the companies that we refer to.

the intensity and ratio images (intensity modulated ratio), but in any case the [Ca²⁺]_ER values cannot be accurately estimated in the underlying ER regions. Movements of the underlying structure during acquisition also tends to smooth out potential differences in [Ca²⁺]_ER between distinct ER regions. Thus, not surprisingly, most studies performed with this optical configuration have reported only spatially averaged [Ca²⁺]ERvalues[9,10,15–17,26,27].

A significant improvement can be achieved by placing a beam splitter in front of the CCD camera[28]. In this case, the emitted light is split by a dichroic/filter combination and the two spectrally separated images are projected on two different portions of the same CCD chip. Proper alignment of the image splitter ensures perfect registration of the two images, allowing the precise calculation of the ratio values even at the organelle edges. The lack of time delay between the two cyan and yellow images also greatly improve the accuracy of the ratio images, and doubles the time resolution.

Another, more expensive solution is to use two cameras to acquire the cyan and yellow images simultaneously[29].

Using a confocal microscope is the best way to increase the spatial resolution. Confocal YC measurements can be performed with single photon excitation using the 458 nm line of an Argon laser [30], but this line is about∼30 nm

(9)

higher than the peak of the ECFP excitation. More efficient excitation of the YC can be achieved with 442 nm line of a HeCd laser, but a specific laser must be purchased for this purpose. Recent solid-state lasers, which can provide up to 25 mW at 430 nm, might be better suited for this application. YC can also be imaged with two-photon excitation in the wavelength range 770–810 nm [31,32].

Video-rate two-photon imaging of YC provides optimal spatial and temporal resolution, but is unfortunately not readily available to most laboratories. A more affordable solution is offered by the use of a spinning disk confocal system (Nipkow), which can be easily implemented on a standard wide-field microscope equipped with a cooled CCD camera [33]. Such a system equipped with a beam splitter on the emission part provides a simple solution for noise-free confocal FRET imaging. The use of a sensitive, back-illuminated frame-transfer CCD camera allows to increase the time-resolution to ∼10 ratio images/s, allowing to resolve rapid, localized [Ca²⁺]_ERchanges.

3.4. Calibration

Like fura-2, the key parameters required to convert YC ratio values into Ca²⁺values are the apparent K_dof the probe, which is assumed to remain constant between experiments, and the maximal and minimal ratio values, which have to be determined for each individual experiment (R_minand R_max).

The apparent Kd measured in living cells can vary significantly from that obtained in vitro, so a titration calibration is highly recommended when existing probes are expressed in a new cell type or organelle or when new probes are generated by genetic engineering. The complete calibration curve can be obtained by incubating cells sequentially with solutions of known Ca²⁺ concentrations buffered with 5 mM EGTA and 5 mM HEEDTA for Ca²⁺ below 100␮M, and 5 mM citrate for Ca²⁺above, using the calculation described in Ref.[34].

Because of the high pH sensitivity of YC, particular care has to be taken to avoid pH variations during the Ca²⁺calibration. The calibration solutions contain 140 mM KCl, pH 7.2 to mimic the composition of the cytosolic and ER compartments, plus 10␮M monensin and 10␮M nigericin to equilibrate the pH across organelle membranes. Equilibra- tion of [Ca²⁺]ER with ionomycin is problematic, as Ca²⁺ must be clamped both in the cytosol and in the ER. Low concentrations of digitonin (5␮g/ml) can be added to favour Ca²⁺equilibration between the external medium and the cytosol. However, higher concentrations should not be used to preserve the organelle integrity and to avoid the loss of YC in the external medium. After obtaining the calibration curves, the apparentK^∗_dand Hill coefficient (n) can be extracted and used to transform the emission ratio (R) in [Ca²⁺] according to the equation:

[Ca²⁺]=K^∗d

R−Rmin

Rmax−R 1/n

Because of cell-to-cell heterogeneity in cytosolic and ER pH as well as in YC fluorescence levels, R_max and R_min have to be systematically determined at the end of each experiment. This implies that at least 30 min has to be devoted to the calibration, a necessary price to get quantitative [Ca²⁺]_ERreadouts with YC. Unlike fura-2, the R_maxvalue is more problematic to obtain than the R_minvalue, because of the low Ca²⁺ affinity of ER-targeted cameleon. Typi- cally, 20 mM Ca²⁺has to be used to get a reliable R_max, and 20 mM EGTA for R_min. Practically, after the experiment, the solution is exchanged to a high K⁺-buffered one containing 20 mM EGTA, pH 7.2. After about 5 min, the ionophores are added (ionomycin, nigericin and monensin) and the cells allowed to equilibrate for ∼20 min to get the Rmin. Then, the solution is switched to 20 mM Ca²⁺, and after 5 min, the ionophores are added again. R_maxis obtained after about 15–20 min of equilibration.

Although this tedious procedure allows to obtain reliable R_max and R_min values, it provides only spatially-averaged values. This is because, at high Ca²⁺concentrations, the ER loses its fine reticular architecture and condenses around the nucleus. Because the ER does not retain its shape during the calibration procedure, a pixel-to-pixel calibration map cannot be obtained. For this reason, it is important to select cells with homogenous ER staining, to maximize the like- lihood that the YC probes will behave similarly in distinct ER regions.

4. Understanding ER Ca²⁺ homeostasis using cameleons

Although the limitations described above have somewhat limited the potential use of the cameleons, several groups have successfully used these probes to obtain new information on ER Ca²⁺homeostasis. Measurements with ER-targeted YC indicated unambiguously that resting [Ca²⁺]_ER values ranged from 250 to 600␮M [9,10,12,15–17,26,35].

These values are in reasonable agreement with earlier measurements using fluorescent dyes or aequorin [4–6,36], and indicate a high degree of cell-to-cell heterogeneity.

Time-resolved [Ca²⁺]ER measurements with YC revealed that the Ca²⁺ turnover across the ER membrane was surprisingly high, both at rest and during agonist stimulation [12,15,17,35]. A significant ER Ca²⁺“leak” was unmasked by inhibition of SERCA ATPases, indicating that the ER has a high passive permeability to Ca²⁺. The leaky nature of the ER might stem from the dual role of this organelle as a protein factory and Ca²⁺ store, as the “leak” pathway appears to be sensitive to puromycin[37]. As a result of the high Ca²⁺leak, an equally high active pumping of Ca²⁺is required to maintain the high resting [Ca²⁺]_ERlevels. Cells thus spend a substantial amount of energy in a futile cycle of Ca²⁺ because the ER is not a “tight” organelle.

Active Ca²⁺ pumping was also prominent during agonist stimulation. In HEK-293 expressing the TRH receptor,

(10)

TRH decreased [Ca²⁺]_ER from 500 to 100␮M in the ab- sence of thapsigargin, and to 5–10␮M in the presence of thapsigargin[15]. This indicates that the ER is continually refilled during agonist stimulation. Accordingly, substantial Ca²⁺ pumping activity was detected when Ca²⁺ was readded to HEK-293 cells stimulated with InsP₃-releasing agonists [35]. Active ER refilling can also be detected as oscillations in [Ca²⁺]_ER which can be occasionally observed during stimulation of HeLa cells with histamine (ND, unpublished observations). The contribution of specific SERCA isoforms to ER refilling was assessed in the insulin-secreting cell line MIN6 using ER-targeted YC. The resting [Ca²⁺]ER decreased from 250 to 100␮M upon mi- croinjection of antisense oligonucleotides directed against the SERCA2b, but not the SERCA3 isoform, implicating the ubiquitous SERCA2b as the major ER Ca²⁺-ATPase in this cell type[17].

YC measurements have allowed to study the link between the ER and plasma membrane channels. The link between [Ca²⁺]_ER and store-operated Ca²⁺ influx has been studied in oocytes and in HEK cells overexpressing the ER-resident Ca²⁺-binding protein calreticulin[16,35]. In both systems, store-operated Ca²⁺influx was found to correlate strictly to [Ca²⁺]_ERlevels, whereas earlier measurements with fura-2 suggested that CCE was inhibited in calreticulin overex- pressers despite full depletion of stores[38]. The strict cor- relation between [Ca²⁺]ER and store-operated Ca²⁺ entry was confirmed by imposing acute changes in [Ca²⁺]ERwith TPEN, a low-affinity membrane-permeant Ca²⁺ chelator [35]. By combining YC measurements with electrophysio- logical recordings in endothelial cells, Frieden et al. have studied the interactions between the ER and K⁺ channels [39]. ER regions close to the plasma membrane were shown to generate local microdomains of high Ca²⁺near the mouth of K⁺ channels, allowing moderate concentrations of agonists to activate endothelial K⁺ channels and to amplify Ca²⁺signals.

The ER-targeted YC have been used to study the link between [Ca²⁺]_ERand apoptosis. Cells expressing the anti- apoptotic protein Bcl-2 had decreased [Ca²⁺]_ERlevels and were more resistant to apoptosis [26], consistent with results obtained with aequorin[40]. Conversely, in cells overexpressing calreticulin, the increased total ER Ca²⁺content correlated with increased [Ca²⁺]ER levels and with an increased susceptibility to apoptotic stimuli [35,41]. In con- trast, cells lacking calreticulin had normal [Ca²⁺]_ERlevels, but a decreased total Ca²⁺content and were more resistant to apoptosis[42]. The induction of the apoptotic cell death program thus appear to correlate with the calcium content of the ER.

YC have also been used to study the interactions between the ER and mitochondria. By measuring the kinetics of [Ca²⁺]_ERchanges in ER regions close or far from mitochondria, mitochondria were shown to favour the local refilling of neighboring ER regions[12]. The presence of mitochondria thus determines the extent of depletion of ER regions,

by enabling the local recycling of Ca²⁺ to vicinal regions of the ER. This cycling of Ca²⁺ between the ER and mitochondria might have important implications, by coupling the function of mitochondria to the ER Ca²⁺content.

In conclusion, cameleons are opening a new window on the cellular aspects of Ca²⁺ signaling. Important information has already been obtained with the first generations of cameleon probes regarding the local dynamics of Ca²⁺ signals inside the ER, the regulation of ER Ca²⁺ fluxes, the interactions between the ER, mitochondria, and plasma membrane channels, and the role of the ER in normal and pathological conditions. The ongoing development of new GFP-based Ca²⁺ indicators coupled with improvements in imaging technology will undoubtedly reveal new and exciting features of the subcellular aspects of Ca²⁺ signals and allow a better understanding of the ER Ca²⁺homeostasis.

Acknowledgements

We thank Dr. O. Hartley for critical reading of the manuscript. ND is a fellow from the Prof. Dr. Max Cloëtta Foundation and his research is supported by operating Grant No. 31-68317.02 from the Swiss National Science Foundation.

References

[1] A.M. Hofer, T.E. Machen, Technique for in situ measurement of calcium in intracellular inositol 1,4,5-trisphosphate-sensitive stores using the fluorescent indicator mag-fura-2, Proc. Natl. Acad. Sci.

U.S.A. 90 (1993) 2598–2602.

[2] K. Hirose, M. Iino, Heterogeneity of channel density in inositol-1,4,5- trisphosphate-sensitive Ca²⁺ stores [see comments], Nature 372 (1994) 791–794.

[3] A. Tse, F.W. Tse, B. Hille, Calcium homeostasis in identified rat gonadotrophs, J. Physiol. (Lond.) 477 (1994) 511–525.

[4] M. Montero, M. Brini, R. Marsault, J. Alvarez, R. Sitia, T. Pozzan, R. Rizzuto, Monitoring dynamic changes in free Ca²⁺concentration in the endoplasmic reticulum of intact cells, EMBO J. 14 (1995) 5467–5475.

[5] D. Button, A. Eidsath, Aequorin targeted to the endoplasmic reticulum reveals heterogeneity in luminal Ca⁺⁺concentration and reports agonist- or IP3-induced release of Ca⁺⁺, Mol. Biol. Cell 7 (1996) 419–434.

[6] M.J. Barrero, M. Montero, J. Alvarez, Dynamics of [Ca²⁺] in the endoplasmic reticulum and cytoplasm of intact HeLa cells. A com- parative study, J. Biol. Chem. 272 (1997) 27694–27699.

[7] M.T. Alonso, M.J. Barrero, P. Michelena, E. Carnicero, I.

Cuchillo, A.G. Garcia, J. Garcia-Sancho, M. Montero, J. Alvarez, Ca²⁺-induced Ca²⁺release in chromaffin cells seen from inside the ER with targeted aequorin, J. Cell Biol. 144 (1999) 241–254.

[8] P. Pinton, D. Ferrari, E. Rapizzi, F.D. Di Virgilio, T. Pozzan, R.

Rizzuto, The Ca²⁺concentration of the endoplasmic reticulum is a key determinant of ceramide-induced apoptosis: significance for the molecular mechanism of Bcl-2 action, EMBO J. 20 (2001) 2690–

2701.

[9] A. Miyawaki, J. Llopis, R. Heim, J.M. McCaffery, J.A. Adams, M.

Ikura, R.Y. Tsien, Fluorescent indicators for Ca²⁺ based on green fluorescent proteins and calmodulin, Nature 388 (1997) 882–887.

(11)

[10] A. Miyawaki, O. Griesbeck, R. Heim, R.Y. Tsien, Dynamic and quantitative Ca²⁺ measurements using improved cameleons, Proc.

Natl. Acad. Sci. U.S.A. 96 (1999) 2135–2140.

[11] V.A. Romoser, P.M. Hinkle, A. Persechini, Detection in living cells of Ca²⁺-dependent changes in the fluorescence emission of an indicator composed of two green fluorescent protein variants linked by a calmodulin-binding sequence. A new class of fluorescent indicators, J. Biol. Chem. 272 (1997) 13270–13274.

[12] S. Arnaudeau, W.L. Kelley, J.V. Walsh Jr., N. Demaurex, Mito- chondria recycle Ca(2+) to the endoplasmic reticulum and prevent the depletion of neighboring endoplasmic reticulum regions, J. Biol.

Chem. 276 (2001) 29430–29439.

[13] J.H. Kim, L. Johannes, B. Goud, C. Antony, C.A. Lingwood, R.

Daneman, S. Grinstein, Noninvasive measurement of the pH of the endoplasmic reticulum at rest and during calcium release, Proc. Natl.

Acad. Sci. U.S.A. 95 (1998) 2997–3002.

[14] O. Griesbeck, G.S. Baird, R.E. Campbell, D.A. Zacharias, R.Y.

Tsien, Reducing the environmental sensitivity of yellow fluorescent protein. Mechanism and applications, J. Biol. Chem. 276 (2001) 29188–29194.

[15] R. Yu, P.M. Hinkle, Rapid turnover of calcium in the endoplasmic reticulum during signaling. Studies with cameleon calcium indicators, J. Biol. Chem. 275 (2000) 23648–23653.

[16] W. Xu, F.J. Longo, M.R. Wintermantel, X. Jiang, R.A. Clark, S.

DeLisle, Calreticulin modulates capacitative Ca²⁺influx by control- ling the extent of inositol 1,4,5-trisphosphate-induced Ca²⁺ store depletion, J. Biol. Chem. 275 (2000) 36676–36682.

[17] A. Varadi, G.A. Rutter, Dynamic imaging of endoplasmic reticulum Ca(2+) concentration in insulin-secreting MIN6 cells using recombinant targeted cameleons: roles of sarco(endo)plasmic reticulum Ca(2+)-ATPase (SERCA)-2 and ryanodine receptors, Diabetes 51 (Suppl. 1) (2002) S190–S201.

[18] K. Truong, A. Sawano, H. Mizuno, H. Hama, K.I. Tong, T.K. Mal, A. Miyawaki, M. Ikura, FRET-based in vivo Ca²⁺ imaging by a new calmodulin-GFP fusion molecule, Nat. Struct. Biol. 8 (2001) 1069–1073.

[19] T. Nagai, A. Sawano, E.S. Park, A. Miyawaki, Circularly permuted green fluorescent proteins engineered to sense Ca²⁺, Proc. Natl.

Acad. Sci. U.S.A. 98 (2001) 3197–3202.

[20] T. Nagai, K. Ibata, E.S. Park, M. Kubota, K. Mikoshiba, A. Miyawaki, A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications, Nat. Biotechnol. 20 (2002) 87–90.

[21] R.E. Campbell, O. Tour, A.E. Palmer, P.A. Steinbach, G.S. Baird, D.A. Zacharias, R.Y. Tsien, A monomeric red fluorescent protein, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 7877–7882.

[22] R. Zufferey, T. Dull, R.J. Mandel, A. Bukovsky, D. Quiroz, L.

Naldini, D. Trono, Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery, J. Virol. 72 (1998) 9873–9880.

[23] M. Ormo, A.B. Cubitt, K. Kallio, L.A. Gross, R.Y. Tsien, S.J. Rem- ington, Crystal structure of the Aequorea victoria green fluorescent protein, Science 273 (1996) 1392–1395.

[24] R.Y. Tsien, The green fluorescent protein, Annu. Rev. Biochem. 67 (1998) 509–544.

[25] A. Miyawaki, R.Y. Tsien, Monitoring protein conformations and interactions by fluorescence resonance energy transfer between mutants of green fluorescent protein, Methods Enzymol. 327 (2000) 472–500.

[26] R. Foyouzi-Youssefi, S. Arnaudeau, C. Borner, W.L. Kelley, J.

Tschopp, D.P. Lew, N. Demaurex, K.H. Krause, Bcl-2 decreases the free Ca²⁺concentration within the endoplasmic reticulum, Proc.

Natl. Acad. Sci. U.S.A. 97 (2000) 5723–5728.

[27] M. Jaconi, C. Bony, S.M. Richards, A. Terzic, S. Arnaudeau, G.

Vassort, M. Puceat, Inositol 1,4,5-trisphosphate directs Ca(2+) flow between mitochondria and the endoplasmic/sarcoplasmic reticulum:

a role in regulating cardiac autonomic Ca(2+) spiking, Mol. Biol.

Cell 11 (2000) 1845–1858.

[28] R. Kerr, V. Lev-Ram, G. Baird, P. Vincent, R.Y. Tsien, W.R. Schafer, Optical imaging of calcium transients in neurons and pharyngeal muscle of C. elegans, Neuron 26 (2000) 583–594.

[29] D.F. Reiff, P.R. Thiel, C.M. Schuster, Differential regulation of ac- tive zone density during long-term strengthening of Drosophila neu- romuscular junctions, J. Neurosci. 22 (2002) 9399–9409.

[30] M. Isshiki, Y.S. Ying, T. Fujita, R.G. Anderson, A molecular sensor detects signal transduction from caveolae in living cells, J. Biol.

Chem. 12 (2002) 12.

[31] G.Y. Fan, H. Fujisaki, A. Miyawaki, R.K. Tsay, R.Y. Tsien, M.H.

Ellisman, Video-rate scanning two-photon excitation fluorescence microscopy and ratio imaging with cameleons, Biophys. J. 76 (1999) 2412–2420.

[32] N. Takahashi, T. Nemoto, R. Kimura, A. Tachikawa, A. Miwa, H.

Okado, Y. Miyashita, M. Iino, T. Kadowaki, H. Kasai, Two-photon excitation imaging of pancreatic islets with various fluorescent probes, Diabetes 51 (Suppl. 1) (2002) S25–S28.

[33] C. Genka, H. Ishida, K. Ichimori, Y. Hirota, T. Tanaami, H.

Nakazawa, Visualization of biphasic Ca²⁺diffusion from cytosol to nucleus in contracting adult rat cardiac myocytes with an ultra-fast confocal imaging system, Cell Calcium 25 (1999) 199–208.

[34] D.M. Bers, C.W. Patton, R. Nuccitelli, A practical guide to the preparation of Ca²⁺buffers, Methods Cell Biol. 40 (1994) 3–29.

[35] S. Arnaudeau, M. Frieden, K. Nakamura, C. Castelbou, M. Michalak, N. Demaurex, Calreticulin differentially modulates calcium uptake and release in the endoplasmic reticulum and mitochondria, J. Biol.

Chem. 24 (2002) 24.

[36] M. Montero, J. Alvarez, W.J. Scheenen, R. Rizzuto, J. Meldolesi, T.

Pozzan, Ca²⁺homeostasis in the endoplasmic reticulum: coexistence of high and low [Ca²⁺] subcompartments in intact HeLa cells, J.

Cell Biol. 139 (1997) 601–611.

[37] R.B. Lomax, C. Camello, F. Van Coppenolle, O.H. Petersen, A.V.

Tepikin, Basal and physiological Ca(2+) leak from the endoplasmic reticulum of pancreatic acinar cells. Second messenger-activated channels and translocons, J. Biol. Chem. 277 (2002) 26479–26485.

[38] L. Mery, N. Mesaeli, M. Michalak, M. Opas, D.P. Lew, K.H. Krause, Overexpression of calreticulin increases intracellular Ca²⁺ storage and decreases store-operated Ca²⁺influx, J. Biol. Chem. 271 (1996) 9332–9339.

[39] M. Frieden, R. Malli, M. Samardzija, N. Demaurex, W.F. Graier, Subplasmalemmal endoplasmic reticulum controls K(Ca) channel activity upon stimulation with a moderate histamine concentration in a human umbilical vein endothelial cell line, J. Physiol. 540 (2002) 73–84.

[40] P. Pinton, D. Ferrari, P. Magalhaes, K. Schulze-Osthoff, F. Di Virgilio, T. Pozzan, R. Rizzuto, Reduced loading of intracellular Ca(2+) stores and downregulation of capacitative Ca(2+) influx in Bcl-2-overexpressing cells, J. Cell Biol. 148 (2000) 857–862.

[41] K. Nakamura, E. Bossy-Wetzel, K. Burns, M.P. Fadel, M. Lozyk, I.S. Goping, M. Opas, R.C. Bleackley, D.R. Green, M. Michalak, Changes in endoplasmic reticulum luminal environment affect cell sensitivity to apoptosis, J. Cell Biol. 150 (2000) 731–740.

[42] K. Nakamura, A. Zuppini, S. Arnaudeau, J. Lynch, I. Ahsan, R.

Krause, S. Papp, H. De Smedt, J.B. Parys, W. Muller-Esterl, D.P.

Lew, K.H. Krause, N. Demaurex, M. Opas, M. Michalak, Functional specialization of calreticulin domains, J. Cell Biol. 154 (2001) 961–

972.

[43] A.L. Facanha, H. Appelgren, M. Tabish, L. Okorokov, K. Ekwall, The endoplasmic reticulum cation P-type ATPase Cta4p is required for control of cell shape and microtubule dynamics, J. Cell Biol.

157 (2002) 1029–1039.

[44] E. Emmanouilidou, A.G. Teschemacher, A.E. Pouli, L.I. Nicholls, E.P. Seward, G.A. Rutter, Imaging Ca²⁺ concentration changes at the secretory vesicle surface with a recombinant targeted cameleon, Curr. Biol. 9 (1999) 915–918.

[45] D.N. Bowser, S. Petrou, R.G. Panchal, M.L. Smart, D.A. Williams, Release of mitochondrial Ca²⁺ via the permeability transition ac-

(12)

tivates endoplasmic reticulum Ca²⁺ uptake, FASEB J. 16 (2002) 1105–1107.

[46] G.J. Allen, J.M. Kwak, S.P. Chu, J. Llopis, R.Y. Tsien, J.F. Harper, J.I.

Schroeder, Cameleon calcium indicator reports cytoplasmic calcium dynamics in Arabidopsis guard cells, Plant J. 19 (1999) 735–747.

[47] G.J. Allen, S.P. Chu, K. Schumacher, C.T. Shimazaki, D. Vafeados, A.

Kemper, S.D. Hawke, G. Tallman, R.Y. Tsien, J.F. Harper, J. Chory,

J.I. Schroeder, Alteration of stimulus-specific guard cell calcium oscillations and stomatal closing in Arabidopsis det3 mutant, Science 289 (2000) 2338–2342.

[48] H. Mizuno, A. Sawano, P. Eli, H. Hama, A. Miyawaki, Red fluores- cent protein from Discosoma as a fusion tag and a partner for flu- orescence resonance energy transfer, Biochemistry 40 (2001) 2502–

2510.

Measurements of the free luminal ER Ca(2+) concentration with targeted &quot;cameleon&quot; fluorescent proteins

Article

Reference